If you remember one thing about the Korean stock market, make it this: the KOSPI is not a diversified portrait of the Korean economy – it is, to a first approximation, a leveraged bet on the memory-semiconductor cycle, because a single company, Samsung Electronics, has historically accounted for roughly one-fifth of the index’s total market value. Add SK Hynix, the other memory giant, and the top two names have at times approached a third of the benchmark. When memory-chip prices turn, the index turns, and no amount of strength in banks, autos, or cosmetics reliably offsets it.
This question resurfaces every time semiconductors sell off. On July 24, 2026, the KOSPI fell 5.7% to close near 6,690 – back below the 7,000 level – in a session led lower by Samsung Electronics and SK Hynix, which fell in unison after a drop in the Philadelphia Semiconductor Index and renewed worries about the profitability of AI capital spending. Foreign investors led the selling. Readers asked the same question they ask every cycle: how can two stocks drag down an index of more than 800 companies? The answer is structural, not situational – so the rest of this guide is written to be just as true a year from now.
One Stock, One Index: The Arithmetic of Concentration
The KOSPI (Korea Composite Stock Price Index) is a capitalization-weighted index covering essentially all common shares listed on the main board of the Korea Exchange (KRX). Capitalization weighting means each company’s influence on the index equals its share of total listed market value – siga chongaek (market capitalization) – with no cap and no equal-weighting adjustment in the headline composite.
Samsung Electronics is, by a wide margin, Korea’s largest listed company. Historically its common shares alone have represented roughly 15-20% of KOSPI market value, and the figure rises further if you include its separately listed preferred shares – useonju (preferred stock, which carries higher dividends but no voting rights). SK Hynix, the world’s other major memory-chip pure-play, has often ranked second. The practical consequence:
- A 3% move in Samsung Electronics alone can move the entire index by roughly half a percent, all else equal.
- Because Samsung Electronics and SK Hynix are exposed to the same demand driver – memory-chip prices – they tend to move together, so their combined effect compounds rather than diversifies.
- The 800-plus smaller constituents matter individually, but collectively they are the tail, not the dog.
Koreans have a word for this: Samsung Electronics is the market’s daejangju (bellwether or “general” stock) – the name whose direction sets the tone for everything else.
How the Cap-Weighted Methodology Works – and Where the Caps Actually Are
It helps to distinguish three related but different indices, because the concentration rules differ:
| Index | Weighting basis | Concentration limit |
|---|---|---|
| KOSPI (composite) | Full market capitalization of all listed common shares | None |
| KOSPI 200 | Free-float-adjusted market cap of 200 large, liquid names | None in the standard version |
| KOSPI 200 capped variants | Free-float-adjusted, with a ceiling (historically around 30%) on any single constituent | Single-name cap applied at rebalancing |
Three points to internalize from that table:
- The headline composite has no cap at all. If Samsung Electronics doubled in value while everything else stood still, its index weight would simply grow. Nothing in the methodology leans against concentration.
- Free-float adjustment barely helps here. The KOSPI 200 excludes strategic holdings from the weight calculation, but Samsung Electronics has a large free float, so its weight remains dominant even after adjustment. In strong semiconductor upcycles it has pressed against the roughly 30% ceiling used in capped variants – which is precisely why those capped versions exist, mainly to keep derivative and fund products within regulatory diversification limits.
- Most index products track the uncapped or lightly capped versions. A plain KOSPI 200 tracker gives you the concentration as-is. If single-name exposure worries you, you must check whether your specific ETF tracks a capped variant – the fund name and prospectus will say so.
The Semiconductor Cycle Is the KOSPI Cycle
Concentration alone would be less dangerous if the giant were a stable utility. It is not. Memory chips – DRAM and NAND – are commodity-like products whose prices swing in multi-year boom-and-bust cycles driven by capacity investment, inventory levels, and end-demand from data centers, smartphones, and PCs. Samsung Electronics and SK Hynix earnings are highly geared to these prices, so the KOSPI inherits the cycle:
- Upcycle: memory prices rise, the two chipmakers’ earnings and multiples expand, their index weight grows, and the KOSPI outruns most regional peers.
- Downcycle: memory prices fall, earnings collapse faster than revenue, and the index underperforms even when domestic consumption and other sectors are fine.
This is why the KOSPI often correlates more tightly with the Philadelphia Semiconductor Index (SOX) than with Korean domestic economic data, and why overnight moves in US chip stocks are the first thing Seoul traders check each morning. Semiconductors are also Korea’s single largest export category, so the same cycle feeds through to the trade balance and, at the margin, the won – meaning an unhedged foreign investor can get hit twice, once on the index and once on the currency.
Worked Example: Reading Foreign Flows Around a Semiconductor Move
Because the largest KOSPI names are also the most liquid, foreign investors – whose net buying is tracked daily as oegugin sunmaesu (foreign net purchases) – express semiconductor views primarily through Samsung Electronics and SK Hynix. Learning to read the daily flow number is therefore a core skill. Here is a real print, step by step.
The figure: foreign net buying across KOSPI and KOSDAQ of +21,359 eok-won on July 23, 2026. One eok is KRW 100 million, so the flow tables you will see on KRX and Naver Finance are quoted in units of KRW 100 million.
- Convert to a usable scale. 21,359 x KRW 100 million = roughly KRW 2.14 trillion of net foreign buying in one session.
- Convert to dollars for intuition. At the USD/KRW rate of about 1,464 (as of July 2026), that is roughly USD 1.5 billion – a genuinely large one-day inflow, not noise.
- Never read one day in isolation. The ten sessions from July 10 to July 24, 2026 ran: -3,228, -16,705, +9,565, +23,031, -13,665, +5,198, +2,952, +26,211, +21,359, -9,374 (all in eok-won). The cumulative total is about +45,000 eok-won, or roughly KRW 4.5 trillion of net buying – yet the window contains single-day swings from -16,705 to +26,211.
- Draw the right conclusion. The trend was accumulation, but the day-to-day sign flipped constantly. A single red day within a positive two-week total is repositioning; five consecutive heavy-selling days is a regime signal. The threshold that matters is persistence, not any one print.
Common mistake to avoid: headlines often report the KOSPI-only foreign flow while data portals show KOSPI+KOSDAQ combined (as in the series above). Always check which universe your number covers before comparing across sources.
Reading the Domestic Cushion: Deposits and Margin Debt
Concentration risk is partly buffered – or amplified – by domestic retail liquidity, which you can monitor with two published aggregates:
- Tujaja yetakgeum (investor deposits): idle cash sitting in brokerage accounts, waiting to be deployed. This stood at roughly KRW 104 trillion as of July 22, 2026. High deposits mean dip-buying firepower exists when foreigners sell the big caps; falling deposits mean the retail cushion is thinning.
- Sinyong yungja (margin loans): money borrowed from brokers to buy stocks, roughly KRW 33 trillion as of July 22, 2026. Elevated margin debt makes semiconductor-led drawdowns sharper, because forced selling of leveraged positions cascades into the same liquid large caps.
The useful read is the ratio and the direction: deposits rising while margin debt is stable suggests a healthy cushion; deposits falling while margin debt climbs is the fragile combination, especially heading into a memory-cycle downturn.
What Concentration Means for Index Investors
- A KOSPI tracker is a semiconductor fund with a Korean wrapper. Before buying, look up the fund’s top-10 holdings and add the weights of Samsung Electronics (common plus preferred, if held) and SK Hynix. If that sum is a quarter or more of the fund, size the position as you would size a sector bet.
- Diversification within Korea is an illusion at the index level. Owning the composite does not spread your risk across 800 businesses; it stacks it on one cycle. If you want broad Korea exposure, pair the index with an equal-weight, capped, or ex-large-cap product – or accept the cycle knowingly.
- Watch the driver, not just the index. Memory-chip spot and contract prices, and the SOX index, lead KOSPI earnings revisions. If you track only the index level, you are reading the output while ignoring the input.
- Mind the currency channel. Because chips dominate exports, semiconductor downcycles often coincide with won weakness. Unhedged foreign holders should evaluate KOSPI products on a currency-adjusted basis, using a dated reference rate (for example, USD/KRW around 1,464 as of July 2026) rather than assuming FX is noise.
- Check which index your ETF actually tracks. “Korea” funds variously track the composite, KOSPI 200, capped variants, or MSCI Korea – each with different Samsung Electronics weights. The single most common retail mistake is assuming these are interchangeable.
FAQ
Why does the exchange not just cap Samsung Electronics in the main index?
The composite is meant to be a neutral measure of total listed market value, not an investment product, so it reflects concentration rather than correcting it. Caps exist in special variants (historically around 30% per name) built mainly so that funds and derivatives can satisfy diversification rules. Investors who want the cap must deliberately choose a product that tracks a capped index.
Is high index concentration unique to Korea?
No. Taiwan’s benchmark is similarly dominated by TSMC, and US indices have grown top-heavy with mega-cap technology names. Korea’s version is distinctive because the dominant exposure is a commodity-like memory cycle, which is more violent than the earnings streams of most mega-caps elsewhere – so the concentration translates into higher index-level cyclicality.
Does buying the KOSDAQ instead solve the concentration problem?
It swaps one concentration for another. The KOSDAQ (Korea’s growth and technology board) is lighter in Samsung Electronics but historically heavy in batteries, biotech, and semiconductor equipment suppliers – many of which are themselves downstream of the same large-cap capex cycle. It diversifies the names, not necessarily the driver.
What is the fastest way to check current index weights myself?
The KRX data portal publishes constituent market caps and index weights, and Naver Finance shows each company’s market value and the daily investor-type flow tables. Dividing Samsung Electronics’ market cap by the total KOSPI market cap – both available on the KRX portal – gives you the live weight in under a minute.
Sources
- KRX data portal – index constituents, market capitalization, investor flows: data.krx.co.kr
- Naver Finance – daily investor-type net buying and company market caps: finance.naver.com
- KOFIA statistics – investor deposits and margin loan balances: freesis.kofia.or.kr
- Bank of Korea ECOS – exchange rates and macro series: ecos.bok.or.kr
Disclaimer: This is an information and analysis publication, not investment advice. See our Methodology for data sources, standards, and our corrections policy.